Skip to main content
European Commission logo print header
Contenu archivé le 2022-12-23

Demonstration of fast parallel architectures for image processing

Objectif



The objective of the project is to implement an image processing system in a joint hardware/software code design cycle. The testbed for the evaluation is the ATLAS detector at the Large Hadron Collider at CERN in Geneva. The real-time performance of the image processor can be tested on-line in supplying continuous data streams from the front-end electronics of the detector input sensor elements. The image processing systems enable the loading of the set of target algorithms and execute the algorithms in real-time in a systolic fashion operating on multi-channel input data streams.

The project is composed of the following actions: firstly, theoretical investigations, including performance evaluation of the algorithms under study, synopsis of the set of target algorithms, evaluation of the space-time complexity of the algorithms (number of I/O channels, number of pixels, bandwidth, search space, etc.); secondly, algorithmical description, including:coding of algorithms in the hardware description language VHDL, behavioural simulations of the algorithms in VHDL, refinement of the algorithms according to functional constraints; thirdly, mapping of algorithms to the targeted FPGA multi-processor system Enable++, including:VHDL synthesis, generating netlists from the VHDL code, partitioning the netlist into several FPGAs, place and route of the netlists, optimisation by rewriting the VHDL code, re-partitioning, re-mapping and re-routing, refinement of the algorithms according to implementational constraints; and finally, Implementation of Enable++ (a multi-FPGA processor system), including architectural design study, printed circuit board layout, equipment and test, construction of high-performance interfaces to the front-end electronics.

At the end of the project a functional system prototype will be available for general evaluation of the acceleration of algorithms to meet the demands of time-critical applications.

Appel à propositions

Data not available

Régime de financement

Data not available

Coordinateur

Universität Mannheim
Contribution de l’UE
Aucune donnée
Adresse
Seminargebäude A5
68131 Mannheim
Allemagne

Voir sur la carte

Coût total
Aucune donnée

Participants (4)